Term
$/MTok (Cost per Million Tokens)
Standard pricing unit for AI APIs — cost in US dollars per one million processed tokens. Listed separately for input, output and sometimes cache.
$/MTok — in more detail
“MTok” stands for million tokens and is the standard pricing unit for AI APIs. A listing like “Claude Sonnet: $3 / $15 per MTok” means: 3 US dollars per million input tokens, 15 US dollars per million output tokens. Rule of thumb for translation: 1,000 tokens correspond to roughly 750 English or 600 German words. One million tokens are roughly half a novel to a full one.
Example / practical context
Sending 100 requests with 5,000 input and 1,000 output tokens each to Claude Sonnet uses 500,000 input and 100,000 output tokens — that’s 0.5 MTok × $3 = $1.50 for input and 0.1 MTok × $15 = $1.50 for output, totalling $3 for the 100 requests. Most provider pricing pages (anthropic.com/pricing, openai.com/api/pricing) list all models side by side in this unit.
Distinction from related concepts
Some providers list prices “per 1k tokens” (1,000 instead of 1,000,000 tokens) — that price is one-thousandth of the MTok price. Local models have no token prices; instead you account for electricity and hardware amortisation as cost per request. With subscription products like ChatGPT Plus you pay a flat monthly fee instead of per token — but only through the chat UI, not the API.
Discover more
AI in Review — August 2026: A Price U-Turn, the Open-Weight Wave and an Enforceable EU AI Act
August 2026 in AI: Sonnet 5 price freeze, Qwen3.8-Max, the open-weight wave, Google AI Mode on Gemini Flash and the enforceable EU AI Act — sorted out.
GlossaryInput Token
Tokens you send to an AI model in an API call — your prompt, the context, attached documents. Billed separately from output tokens and usually much cheaper.
EncyclopediaFunction Calling / Tool Use
How an LLM uses tools: define a tool as a schema, the model picks the function and arguments, the result returns to the chat — the basis of every agent.